View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biophysical Journal Volume 87 August 2004 1165–1172 1165

Lamellar Organization of Pigments in Chlorosomes, the Light Harvesting Complexes of Green Photosynthetic

J. Psˇencˇ´ık,*T.P.Ikonen,y P. Laurinma¨ki,z M. C. Merckel,§ S. J. Butcher,z R. E. Serimaa,y and R. Tumaz *Department of Chemical Physics and Optics, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic; and yDivision of X-ray Physics, Department of Physical Sciences, zInstitute of Biotechnology and Department of Biological and Environmental Science, and §Helsinki Bioenergetics Group, Institute of Biotechnology, University of Helsinki, Helsinki, Finland

ABSTRACT Chlorosomes of green photosynthetic bacteria constitute the most efficient light harvesting complexes found in nature. In addition, the chlorosome is the only known photosynthetic system where the majority of pigments (BChl) is not organized in pigment- complexes but instead is assembled into aggregates. Because of the unusual organization, the chlorosome structure has not been resolved and only models, in which BChl pigments were organized into large rods, were proposed on the basis of freeze-fracture electron microscopy and spectroscopic constraints. We have obtained the first high- resolution images of chlorosomes from the green sulfur bacterium tepidum by cryoelectron microscopy. Cryoelectron microscopy images revealed dense striations ;20 A˚ apart. X-ray scattering from chlorosomes exhibited a feature with the same ;20 A˚ spacing. No evidence for the rod models was obtained. The observed spacing and tilt-series cryoelectron microscopy projections are compatible with a lamellar model, in which BChl molecules aggregate into semicrystalline lateral arrays. The diffraction data further indicate that arrays are built from BChl dimers. The arrays form undulating lamellae, which, in turn, are held together by interdigitated esterifying alcohol tails, , and . The lamellar model is consistent with earlier spectroscopic data and provides insight into chlorosome self-assembly.

INTRODUCTION Photosynthesis is the ultimate source of energy for most sions 1500 3 500 3 200 A˚ ,1A˚ ¼ 0.1 nm) and the large current life forms, including humans. The first step in number of pigment molecules inside. A typical chlorosome utilization of solar energy is photon capture by a light- contains on the order of 105 BChl molecules (Montano et al., harvesting system (antenna). Typically, the antennae are 2003) (BChl c, d, e, depending on the species) in the form of composed of pigment-protein complexes, in which the aggregates. The aggregation modulates the optical properties protein framework determines pigment orientations and of the BChls and results in fast energy transfer rates optical properties, and ensures efficient flow of excitation (Prokhorenko et al., 2000; Psencik et al., 2003, and energy to the photosynthetic reaction center. The only references therein), which are a prerequisite for the light- known exception is the chlorosome. Chlorosomes are large harvesting efficiency. enclosures of BChl aggregates, which are organized by Chlorosomes were first reported in 1963 and their pigment-pigment rather than pigment-protein interactions, structure was subsequently characterized by electron mi- and are attached to the inner side of the cytoplasmic croscopy (Cohen-Bazire et al., 1964). The electron micro- membrane of green photosynthetic bacteria (Blankenship graphs of Cohen-Bazire and co-workers revealed 12–20 A˚ et al., 1995; Frigaard et al., 2003). Two bacterial families, the wide striations, arrayed more or less parallel to the long axis green sulfur and the green nonsulfur bacteria, belong to this of the chlorosome, but no interpretation was given. Later, group and are only distantly related, but use chlorosomes as this observation was overshadowed by freeze-fracture the main light harvesting system. The electron microscopy study (Staehelin et al., 1978, 1980), in are able to survive at the lowest light conditions of all known which micrographs of chlorosome interiors were interpreted photosynthetic organisms (Overmann et al., 1992; Frigaard in terms of rod-like elements with a diameter of 50 A˚ et al., 2003). In effect, the chlorosome is the most efficient (nonsulfur bacteria) or 100 A˚ (sulfur bacteria). This antenna known. The efficiency is in part due to the large size interpretation was reinforced by freeze-fracture (Oelze and of the chlorosome (ellipsoidal particle with typical dimen- Golecki, 1995) and disruption (Wullink and van Bruggen, 1988) studies and became the basis for all subsequent chlorosome models. Perhaps because of the large size and unusual organization, no crystals of chlorosomes or ag- Submitted January 21, 2004, and accepted for publication May 20, 2004. gregated BChls have been obtained, and the chlorosome Address reprint requests to R. Tuma, Institute of Biotechnology, remains the last known light-harvesting complex for which Viikinkaari 1, PL 65, University of Helsinki, FIN-00014, Helsinki, Finland. Tel.: 358-9-19159577; Fax: 358-9-19159930; E-mail: roman.tuma@ no high-resolution structural information is available. helsinki.fi. Several models for the organization of BChl aggregates Abbreviations used: BChl, ; Chl., chlorobium; EM, into rod-like elements were proposed (Holzwarth and cryoelectron microscopy; OD, optical density (cm1); SAXS, small angle Schaffner, 1994; Nozawa et al., 1994; Blankenship et al., x-ray scattering; WAXS, wide angle x-ray scattering. Ó 2004 by the Biophysical Society 0006-3495/04/08/1165/08 $2.00 doi: 10.1529/biophysj.104.040956 1166 Psˇencˇ´ıketal.

1995; van Rossum et al., 2001). These models can be another round of density-gradient centrifugation. The density and the optical classified into two principal groups based on the asymmetric absorption spectrum of the resulting chlorosomes were measured to assure repeating unit: 1), parallel-chain model with BChl monomer sample integrity. The absorption spectra were measured before and after each experiment to ensure that no degradation occurred during exposure to as the building block (Holzwarth and Schaffner, 1994; x rays or during the handling required for EM sample preparation. Balaban et al., 1995; Chiefari et al., 1995; van Rossum et al., 2001); and 2), antiparallel double-chain model with an antiparallel so-called ‘‘piggy-back’’ BChl dimer as a building Electron microscopy block (Smith et al., 1986; Nozawa et al., 1994; Umetsu et al., Small drops (10 ml) of fresh chlorosome solution (OD ;80 per cm at 748 1999; Wang et al., 1999a; Umetsu et al., 2002). The terms nm) were dialyzed for 1 min using Millipore 0.025 mm pore membrane parallel and antiparallel refer to the mutual orientation of the (Millipore, Billerica, MA), against 5 mM Tris, pH 8. Protein-A gold (50 A˚ , Q transition dipoles. In all these models, short-range order Dept. of Cell Biology, University of Utrecht) was added to chlorosome y solutions before dialysis as fiducial markers for tilt experiments. Grid was based on the results of NMR and optical spectroscopy preparation and vitrification was accomplished by the guillotine method of (Smith et al., 1986; Hildebrandt et al., 1991; Holzwarth and Dubochet et al. (1988) in liquid-nitrogen cooled ethane. The vitrified Schaffner, 1994; Nozawa et al., 1994; Balaban et al., 1995; samples were examined in a Tecnai F20 transmission electron microscope Chiefari et al., 1995; Umetsu et al., 1999, 2002; Wang et al., using an Oxford CT3500 cryo-holder (EM Unit, Institute of Biotechnology, 3 1999a; Mizoguchi et al., 2000; van Rossum et al., 2001). University of Helsinki). Micrographs were recorded at 200 kV, 50,000 magnification, 0.8–3.2 mm underfocus, on Kodak S0163 film using low Most notably, the long-range order followed from constrain- dose. Images free from astigmatism and drift were scanned at 7 mm step size ing the models to resemble the rod-like elements of Staehelin (1.4 A˚ per pixel) using a Zeiss Photoscan TD scanner. The defocus of the et al. (1978, 1980). Furthermore, none of these models has scanned micrographs was calculated using the program CTFFIND3 provided insight into chlorosome assembly. (Grigorieff, 1998). Individual chlorosomes were boxed (1000 3 1000 In addition to BChl aggregates, chlorosomes contain BChl pixels) from the original images and subregions containing the fine structure were selected (area of 125 3 125 pixels). The subregions were padded to a, carotenoids, quinones, lipids, and . Lipids pre- 2048 3 2048 array with the average image intensity, and power spectra were sumably form an enveloping monolayer of the chlorosome calculated. and the coupling to the cytoplasmic membrane is achieved via a BChl a-containing protein baseplate (Blankenship et al., 1995; Frigaard et al., 2003). Proteins constitute a minor X-ray scattering component and are thought to reside in the chlorosome Samples for x-ray scattering were prepared by rapid dialysis against 1 mM baseplate and envelope (Blankenship et al., 1995; Frigaard Tris, pH 8, followed by controlled concentration under low vacuum to avoid et al., 2003). complete drying and salt crystallization. The concentrated but fluid sample (OD ;2000 per cm at 748 nm) was loaded into a steel-framed sample cell Here we present the first EM images of intact Chlorobium (thickness 1 mm) sealed with two 13 mm Capton windows. The tepidum chlorosomes embedded in vitreous ice. EM was measurements in the q-range of 0.1–2.5 A˚ 1 (q ¼ 4 3 p sin(q/2)/l, where complemented by solution SAXS and WAXS. Both the q is the scattering angle and l is the wavelength) were made on the in-house high-resolution EM images and SAXS revealed fine internal x-ray scattering apparatus in Division of X-ray Physics, University of structure with spacing of ;20 A˚ , which can be explained by Helsinki. The quasi-monochromatic x rays (Cu Ka) scattered from the sample were detected with a two-dimensional proportional counter. The a simple lamellar organization of pigment molecules. The x-ray transmission of the samples was determined during the scattering data were clearly incompatible with the rod-like model. experiments using a transparent beam stop. The q-scale was calibrated with Based on these results we propose a new lamellar structure of a standard sample of silver behenate. chlorosomes, which yields a model of chlorosome assembly. Measurements at very low-angles were conducted at the synchrotron beamline BW4 in HASYLAB, Hamburg, Germany (Gehrke, 1992). The energy of the x rays was 8.979 keV and two measurement distances of 4 m and 13 m were used. The q-ranges achieved with these two configurations MATERIALS AND METHODS were 0.011–0.10 A˚ 1 and 0.0028–0.033 A˚ 1, respectively. The beam intensity and sample absorption was monitored by two ionization chambers, Chlorosome preparation and characterization placed before and after the sample, and by a PIN diode at the beamstop. A standard sample of rat tail collagen was used to calibrate the q-scale. The Chl. tepidum cultures (kind gift of Prof. G. Hauska, University of In all experiments, the measured data were corrected for the detector Regensburg) were grown for 3 days at 48°C in a modified Pfennig’s medium response and averaged circularly to obtain one-dimensional scattering (Wahlund et al., 1991) under constant illumination (60 W Tungsten lamp, curves. The background scattering from the cuvette and solvent buffer were 25 cm illumination distance). The culture was stored at 4°C (dark) and measured separately. The intensity I of scattering by the sample was harvested by centrifugation (7000 RPM, Sorvall SLA-3000 rotor, 10 min, obtained (in arbitrary units) by subtracting the weighed intensity I of the 4°C). The chlorosomes were isolated by a method of Gerola and Olson b cuvette and the solvent background from the sample measurement I : (1986) with modifications. Cell pellets from 250 ml culture were s resuspended in 5 ml of 50 mM Tris buffer pH 8, containing 2M sodium ¼ ð Þð = Þ ; isothiocyanide (NaSCN) and lysed by three passages through a French I Is 1 f Ts Tb Ib pressure cell at 20,000 psi. Cell debris was removed by centrifugation

(10,000 RPM, Sorvall SS-34 rotor, 10 min, 4°C). The chlorosome- where Ts and Tb are the transmission coefficients of the sample and containing supernatant was loaded onto 10–40% sucrose density gradient background, respectively, and f is the estimated volume fraction of the and centrifuged for 24 h at 220,000 g, 5°C. The chlorosome-containing band chlorosomes. A volume fraction of 0.2 was estimated for the sample used in was recovered, concentrated by centrifugation, and further purified by measurements with synchrotron radiation, whereas the sample used in the

Biophysical Journal 87(2) 1165–1172 Chlorosome Structure 1167 wide angle measurements with the conventional x-ray source was more dimensions 1400–1800 A˚ 3 500 A˚ . The most remarkable concentrated, having a volume fraction of 0.3. feature, which is observed in nearly all chlorosomes, is In addition, a wide-angle x-ray scattering measurement was made in- a striation pattern formed by parallel dark and light stripes house (as described above) from completely dry, flake-like sample of chlorosomes. The scattering background corresponding to an empty sample oriented close to parallel with the long axis of the holder was subtracted. chlorosome. The distance between centers of two neighboring dark (or light) striae is ;20 A˚ . The distance was determined more RESULTS precisely by calculating Fourier transforms of selected regions. A single intense band was usually observed for Electron microscopy the chlorosome region showing pronounced striation (Fig. To retain high-resolution information, chlorosomes were 1). The peak was always found on the axis perpendicular to embedded in vitreous ice and imaged without further the direction of the striation (long axis of the chlorosome). treatment or staining. Fig. 1 shows typical projection images The distance of the peak from the center was found in of chlorosomes obtained by EM. The chlorosomes appear as a narrow range corresponding to spacing of 19.5–21.5 A˚ approximately elliptic objects with ruffled edges and exhibit with a mean value of 20.5 A˚ .

FIGURE 1 EM analysis of chlorosomes. Image of four representative chlorosomes embedded in vitreous ice. The bottom panels show power spectra of the boxed areas and the corresponding striation spacings. Bar, 500 A˚ . The defocus value for panels a, b, and d was 2.5 mm and 2.55 mm for panel c.

Biophysical Journal 87(2) 1165–1172 1168 Psˇencˇ´ıketal.

FIGURE 2 (Upper panels) Tilted series images (15, 0, and 30°) of a representative chlorosome. Bar, 500 A˚ . The defocus value was 1.6 mm. The tilt axis is oriented almost parallel to the striae and intersects through the middle of the chlorosome. Lower panels show the power spectra computed from the boxed regions (shown as a square in each upper panel), which were selected to correspond to the same volume of the chlorosome throughout the tilt.

To further characterize the origin of the striation, tilt series would be expected for hexagonally arranged rod-like were collected. Three tilt angles (15, 0, and 30°) were elements with diameter of 50 to 100 A˚ (cf. the solid and selected to determine a plausible pigment packing. For dashed lines in Fig. 3). Moreover, the scattering pattern does example, hexagonally packed pigments would yield projec- not exhibit features typical for a dilute solution of rods tions in which the spacing would strongly depend on the tilt (Gandini et al., 2003). Instead, the scattering curve contains angle. Striae with similar spacing were observed under a prominent scattering peak at q ¼ 0.30 A˚ 1 which different tilts, often within the same region of the corresponds to a spacing of 20.9 A˚ . Additional maxima are chlorosome. An example is shown in Fig. 2 of striation observed at 0.54, 0.67, and 0.76 A˚ 1 (Bragg distances 11.7, observed under 15 and 30° tilts. In other areas, striation 9.4, and 8.2 A˚ , respectively). The four observed diffraction was observed under both 0° and 30° tilts (not shown), which maxima could not yield a unique lattice for pigment made a hexagonal arrangement of pigments unlikely. A arrangement. However, a preliminary assignment, which is splitting of the diffraction peak in the power spectrum (Fig. compatible with both EM and diffraction is a monoclinic 2, 130°, bottom right) indicates that two domains with the lattice (a ¼ 9.6 A˚ , b ¼ 12.0 A˚ , c ¼ 20.9 A˚ , and g ¼ 77.5°, same spacing but slightly different orientation coexist within Fig. 3, inset). Additionally, a broad feature at wide angles the projected volume. with a maximum at 1.4 A˚ 1 (spacing 4.5 A˚ ) was observed (Fig. 3), which reflects short-range order, e.g., stacking of chlorin rings and intramolecular electron density correlation. X-ray scattering The measurement of the dry chlorosome sample (Fig. 3, To determine the internal arrangement of the pigments, e.g., inset) contains the same structural features as the hydrated size and spacing of the putative rods, x-ray scattering from sample, demonstrating that water does not play a significant the chlorosome solution was obtained (Fig. 3). The initial role in the internal structure of the chlorosome. The absence portion of the scattering curve (q , 0.05) reflects the of sharp diffraction maxima in the wide angles of both the ellipsoidal shape of chlorosomes. Notably, no peak was hydrated and the nonhydrated samples indicates that there is observed in the range of q values 0.06–0.15 A˚ 1, which no rigid long-range order in the pigment packing.

Biophysical Journal 87(2) 1165–1172 Chlorosome Structure 1169

FIGURE 3 X-ray scattering curves measured from a concentrated chlorosome solution combined from four individual measurements. The dashed curve shows scattering calculated from an analytical model containing 1000 A˚ long hollow cylinders of inner radius of 40 A˚ and outer radius of 50 A˚ in a 5 3 2 hexagonal lattice with a lattice constant of 100 A˚ (Prokhorenko et al., 2000), plus an exponentially decaying background. (Inset) X-ray scattering from a dry chlorosome sample. The positions of the diffraction maxima and corresponding monoclinic lattice indices are indicated by arrows.

DISCUSSION Cryoelectron imaging and x-ray scattering reveal internal structure of the chlorosome An important question is whether the striation observed in the EM images originates from the projection of BChl c aggregates inside the chlorosome or from the surface features (the baseplate or envelope). The latter possibility is unlikely, however, given the following arguments. The EM image is a projection of the whole chlorosome, in which the surface, namely the envelope and the baseplate, provides less contrast compared to the internal volume. The same argument holds for scattering power in SAXS. Furthermore, the ;20 A˚ striation is seen in the same area under tilt angles of 15° and 30° (i.e., 45° difference) which would be impossible for a surface feature with the same spacing (Fig. 2). Finally, the FIGURE 4 Schematic model of the BChl aggregates in the chlorosome. strongly interacting and ordered chlorin rings are likely to (A) Arrangement of lamellae inside the chlorosome. Each undulated plane contribute to the observed diffraction pattern. Thus, we (thick green line) extends in the long axis of the chlorosome (z), and through conclude that the ;20 A˚ spacing arises from the volume of the height of the chlorosome (y). Planes are arranged into lamellae the chlorosome rather than from its surface and reflects the throughout the chlorosome (x). Only a few lamellae are depicted for clarity. organization of the BChl aggregates. Interestingly, a similar Model of one plane of BChl aggregate with parallel (B) or antiparallel (C) ˚ pigment configuration. Coordinate system as in panel A. The monoclinic 12–20 A striation was reported by Cohen-Bazire et al. (1964) unit cell projection and lattice constants are shown in red. (D) Top view of for chlorosomes in sectioned cells and was observed within chlorin (green) planes (antiparallel configuration) associated via interdigi- the chlorosome interior, remote from the baseplate. tated esterifying alcohol tails (black). An underlying layer of BChl molecules is shown dotted whereas carotenoids (orange) are interspersed between the alcohol tails. Chlorosome pigments cannot be arranged in rod-like elements Both experimental methods (SAXS and EM) show that the dominant spacing in the chlorosome is ;20 A˚ . This value is

Biophysical Journal 87(2) 1165–1172 1170 Psˇencˇ´ıketal. clearly incompatible with the spacing of 86.6 A˚ expected for coordination and/or chlorin ring stacking (Fig. 4). In this a hexagonal lattice of rods with 100 A˚ diameter (86.6 A˚ ¼ arrangement the esterifying alcohol tails stick out of the 100 A˚ /cos(30°)) as proposed in all current models. Notably, plane (Fig. 4, B and C) (Holzwarth and Schaffner, 1994; absence of maxima between ;40–100 A˚ in the measured Nozawa et al., 1994) as hydrophobic surfaces for lamellar SAXS (Fig. 3) and EM makes the presence of 50–100 A˚ assembly (Fig. 4 D). diameter rods in any arrangement unlikely. The monoclinic lattice that we propose based on x-ray In principle, the SAXS data could be explained by scattering could provide clues about the pigment arrange- pigments arranged in rods (20 or 24.1 A˚ diameter), which are ment within the planes. The lattice dimensions are most organized on an oblique or hexagonal, two-dimensional, compatible with a dimer in the unit cell as shown in Fig. 4 C. lattice, respectively. The latter possibility is not compatible Indeed, the antiparallel dimer is stable in solution (Smith with EM results. Attempts to build plausible molecular et al., 1986; Wang et al., 1999a,b; Umetsu et al., 2002) and models with rods of 20 A˚ diameter, which would satisfy has been suggested as the building block of larger aggregates spectroscopic results and hydrogen bonding patterns, were (Brune et al., 1988; Nozawa et al., 1994; Umetsu et al., 1999; unsuccessful due to the following reason: close packing Wang et al., 1999a; Mizoguchi et al., 2000). Assuming the required interdigitation of the rods and led to unrealistically dimer model and the proposed lattice arrangement (Fig. 4 C) high densities (.1.9 g/cm3). the stacking distance between chlorin rings is 3.3 A˚ whereas the Mg-Mg distance between BChls along the z axis is 16.9 A˚ . An alternative lattice orientation with the short diagonal Lamellar model along the z axis gives the stacking distance as 4.2 A˚ , whereas As all of our results indicated that the rod model was not the Mg-Mg distance is 13.6 A˚ . The broad WAXS feature feasible, we searched for other plausible arrangements. EM (q-range 1–2.5 A˚ 1 in Fig. 3, inset), notably the absence of images (Fig. 1) suggested that pigments could be arranged in a prominent peak corresponding to a fixed stacking distance parallel planes, i.e., lamellae parallel to the long axis of the between chlorin rings, suggests considerable disorder within chlorosome. Indeed, the most prominent SAXS feature at the planes. Because of the disorder, the above distances q ¼ 0.30 A˚ 1 (Fig. 3) is consistent with the lamellar model represent limiting cases and the actual values will fluctuate with ;20 A˚ spacing between planes. within the range of 3.3–4.2 A˚ and 13.6–16.9 A˚ , respectively. The EM tilt series gave further information about the These estimates compare favorably with the range of values organization of lamellae within the chlorosome. If the obtained for the antiparallel dimer model from NMR lamellae maintained planarity throughout the chlorosome, constraints (stacking distance 3.2–3.4 A˚ (Wang et al., the striation should disappear as the chlorosome is tilted. 1999a), Mg-Mg distance between BChls along the z axis However, this was not the case and the striae were observed 15.25–15.5 A˚ (Nozawa et al., 1994; Mizoguchi et al., 2000)). under different tilts often from the same region of the The proposed model yields a density of ;1.15 g/ml (without chlorosome (Fig. 2). Thus, the lamellae seem to undulate carotenoids) or 1.21 (with 6% w/w carotenoids, Borrego in the direction perpendicular to the long axis of the et al., 1999), which is similar to the measured density of chlorosome. The width of the 20.9 A˚ SAXS peak also chlorosomes (1.16 6 0.05 g/ml). However, the monoclinic indicates that the order persists only over 60–80 A˚ distance lattice constitutes only the first approximation and de- and suggests considerable disorder. Fig. 4 A summarizes the lineation of detailed molecular arrangements within the lamellar model of the chlorosome structure. planes will have to await further experiments. Thus, our results are compatible with an antiparallel dimer arrangement similar to those based on NMR. The antiparallel The lamellar model is consistent with structural chain model also nicely explains results of Stark spectros- and spectroscopic results copy (Frese et al., 1997), which demonstrated a lack of It remained to be seen whether the lamellar arrangement of permanent dipole moment difference between the ground- pigments could be achieved using the well-established short- and excited states in chlorosomes (note that individual BChl range interaction patterns of BChl molecules. Two possible c possesses such dipole moment difference). In the configurations of BChl molecules have been suggested for antiparallel dimer the dipole moments compensate each aggregates on the basis of NMR experiments: the parallel- other, which is not the case for the parallel arrangement. chain model (Balaban et al., 1995) (Fig. 4 B) and the The EM and SAXS data indicate that the aggregates are antiparallel-chain model (Nozawa et al., 1994) (Fig. 4 C). significantly disordered by undulation. The disorder and The interactions satisfied by these models are coordination of undulation in the direction perpendicular to the z axis can be Mg atoms by the C31-OH of the stacked neighbor and rationalized within the context of the antiparallel model. In a hydrogen bond between the C31-OH and C131 ¼ O of the this model the interfaces between the dimer chains are not neighbor in the same chain (Fig. 4, B and C). stabilized by hydrogen bonding (Fig. 4 C) leading to weaker We propose that the linear chains as described above can interactions than within the dimer and providing preferable be arranged into planar structures instead of rods via Mg interface for line defects and bending.

Biophysical Journal 87(2) 1165–1172 Chlorosome Structure 1171

Despite the disorder, the lamellar packing produces an between esterifying alcohol chains in the outer part of alteration of the electron-dense, tightly packed, chlorin the chlorosome and hydrophobic tails of monogalactosyl layers with the less dense, -like, interdigitated esterify- diglyceride, which is thought to form the chlorosome ing alcohol interfaces (Fig. 4 D). Such alteration provides envelope. enough contrast to be observed by EM and gives rise to the SAXS pattern. However, the striation would be hard to detect in freeze-fracture electron micrographs and the rod- CONCLUSIONS like elements observed by Staehelin et al. (1978, 1980) most EM and x-ray scattering data rule out the existence of likely originated from the cleavage along the undulated previously proposed rod-like aggregates with 50–100 A˚ (oblique) planes within the lamellae. diameter in chlorosomes. A new lamellar model of pigment EM projection and tilt series show that the lamellae are organization is proposed. The model is consistent with preferentially ordered in the direction of the long chlorosome current and previously published structural and spectro- axis (z) whereas undulation and disorder is prevalent in the scopic constraints and provides insight into the assembly of perpendicular directions (Fig. 4 A). Because the transition the chlorosome. dipole moments of the BChl c molecules in the model are nearly parallel with the z axis of the chlorosome (Fig. 4 D) The authors express their gratitude to Prof. G. Hauska (University of the observed strong order along this axis will produce strong Regensburg) for providing the Chl. tepidum strain and to Dr. F. Vacha linear dichroism (Blankenship et al., 1995; Frese et al., (University of South Bohemia) for growing the bacteria. HASYLAB is thanked for providing beam time. 1997). The disorder observed in SAXS and expected to originate from undulations along the y axis may well explain This study was supported by the Academy of Finland (projects 206926, R.T.; 208661, S.J.B.; 205138, J.P.; and 172622, R.E.S.; and the Finnish the broad NMR resonances previously observed and at- Centre of Excellence Program 2000-2005, R.T and S.J.B.) and by the tributed to two forms of BChl c aggregates in the bilayer rod Czech Science Foundation and Czech Ministry of Education, Youth and model (van Rossum et al., 2001). Sports (contracts 206/02/0942, LN00A141 to J.P.). T.P.I. is supported by the National Graduate School in Informational and Structural Biology.

Chlorosome assembly The lamellar organization presented here provides clues REFERENCES about how self-assembly of the chlorosome may occur. The Balaban, T. S., A. R. Holzwarth, K. Schaffner, G. J. Boender, and H. J. de BChl a molecules and proteins of the crystalline baseplate Groot. 1995. CP-MAS 13C-NMR dipolar correlation spectroscopy of 13C-enriched chlorosomes and isolated bacteriochlorophyll c aggregates form a regular lattice and may serve as the nucleation site for of Chlorobium tepidum: The self organization of pigments is the main lamellar assembly. The rest of the assembly process proceeds structural feature of chlorosomes. Biochemistry. 34:15259–15266. via propagation of the planes utilizing the nonspecific Blankenship, R. E., J. M. Olson, and M. Miller. 1995. Antenna complexes hydrophobic interactions of the tails while maintaining order from green photosynthetic bacteria. In Anoxygenic Photosynthetic via the specific interactions of the chlorin rings. Arrangement Bacteria. R. E. Blankenship, M. T. Madigan, and C. E. Bauer, editors. Kluwer Academic Publisher, Dordrecht, The Netherlands. 399–435. of BChl molecules, in which the alcohol tails extend from both sides of the plane, would permit interdigitation of the Borrego, C. M., P. G. Gerola, M. Miller, and R. P. Cox. 1999. Light intensity effects on pigment composition and organisation in the green hydrophobic tails and facilitate assembly of the planes into sulfur bacterium Chlorobium tepidum. Photosynth. Res. 59:159–166. the lamellae. The antiparallel configuration yields automat- Brune, D. C., G. H. King, and R. E. Blankenship. 1988. Interactions ically such arrangement of the tails (Nozawa et al., 1994) between bacteriochlorophyll c molecules in oligomers and in chloro- (Fig. 4 C) although a parallel configuration with similar somes of green photosynthetic bacteria. In Photosynthetic Light- Harvesting Systems. H. Scheer,and S. Schneider, editors. Walter de properties can be envisioned (not shown). Gruyter, Berlin. 141–151. The model also suggests that hydrophobic carotenoids Chiefari, J., K. Griebenow, F. Fages, N. Griebenow, T. S. Balaban, A. R. may occupy the lipid-like space between the planes Holzwarth, and K. Schaffner. 1995. Models for the pigment organization interacting with the esterifying alcohol chains (Fig. 4 D). in the chlorosomes of photosynthetic bacteria: Diastereoselective control The volume fraction of carotenoids within this layer may of in vivo bacteriochlorophyll cs aggregation. J. Phys. Chem. 99:1357– 1365. then modulate its thickness and the lamellar spacing. Indeed, preliminary SAXS and EM examination of Chl. phaeobac- Cohen-Bazire, G., N. Pfennig, and R. Kunisawa. 1964. The fine structure of green bacteria. J. Cell Biol. 22:207–225. teroides chlorosomes, which possess a larger amount of Dubochet, J., M. Adrian, J. J. Chang, J. C. Homo, J. Lepault, A. W. carotenoids compared to Chl. tepidum, yielded a spacing of McDowall, and P. Schultz. 1988. Cryo-electron microscopy of vitrified 28 A˚ (unpublished data) as opposed to ;20 A˚ obtained for specimens. Q. Rev. Biophys. 21:129–228. Chl. tepidum in this study. This tendency fits well with the Frese, R., U. Oberheide, I. H. M. van Stokkum, R. van Grondelle, M. Foidl, results of x-ray diffraction experiments previously reported J. Oelze, and H. van Amerongen. 1997. The organization of for BChl c aggregates in nonpolar solvents (without bacteriochlorophyll c in chlorosomes from Chloroflexus aurantiacus ; and the structural role of carotenoids and protein—an absorption, linear carotenoids) where the largest observed spacing was 18 dichroism, circular dichroism and Stark spectroscopy study. Photosynth. A˚ (Umetsu et al., 1999). Similar interactions are expected Res. 54:115–126.

Biophysical Journal 87(2) 1165–1172 1172 Psˇencˇ´ıketal.

Frigaard, N. U., A. G. M. Chew, H. Li, J. A. Maresca, and D. A. Bryant. Prokhorenko, V. I., D. B. Steensgaard, and A. R. Holzwarth. 2000. Exciton 2003. Chlorobium tepidum: insights into the structure, physiology, and dynamics in the chlorosomal antennae of the green bacteria Chloroflexus metabolism of a green sulfur bacterium derived from a complete genome aurantiacus and Chlorobium tepidum. Biophys. J. 79:2105–2120. sequence. Photosynth. Res. 78:93–117. Psencik, J., Y. Z. Ma, J. B. Arellano, J. Hala, and T. Gillbro. 2003. Gandini, S. C. M., E. L. Gelamo, R. Itri, and M. Tabak. 2003. Small angle Excitation energy transfer dynamics and excited-state structure in x-ray scattering study of meso-tetrakis (4-Sulfonatophenyl) porphyrin in chlorosomes of Chlorobium phaeobacteroides. Biophys. J. 84:1161– aqueous solution: a self-aggregation model. Biophys. J. 85:1259–1268. 1179. Gehrke, R. 1992. An ultrasmall angle scattering instrument for the doris-III Smith, K. M., F. W. Bobe, D. A. Goff, and R. J. Abraham. 1986. NMR bypass. Rev. Sci. Instrum. 63:455–458. spectra of porphyrins. 28. Detailed solution structure of bacteriochloro- phyllide d dimer. J. Am. Chem. Soc. 108:1111–1120. Gerola, P. D., and J. M. Olson. 1986. A new bacteriochlorophyll a-protein complex associated with the chlorosomes of green sulfur bacteria. Staehelin, L. A., J. R. Golecki, R. C. Fuller, and G. Drews. 1978. Biochim. Biophys. Acta. 848:69–76. Visualization of the supramolecular architecture of chlorosome (Chloro- bium type vesicles) in freeze-fractured cells of Chloroflexus aurantiacus. Grigorieff, N. 1998. Three-dimensional structure of bovine NADH: Arch. Microbiol. 119:269–277. Ubiquinone oxidoreductase (Complex I) at 22 angstrom in ice. J. Mol. Biol. 277:1033–1046. Staehelin, L. A., J. R. Golecki, and G. Drews. 1980. Supramolecular organization of chlorosome (Chlorobium vesicles) and of their membrane Hildebrandt, P., K. Griebenow, A. R. Holzwarth, and K. Schaffner. 1991. attachment site in Chlorobium limicola. Biochim. Biophys. Acta. 589: Resonance Raman spectroscopic evidence for the identity of the 30–45. bacteriochlorophyll c organisation in protein-free and protein-containing Umetsu, M., Z. Y. Wang, J. Zhang, T. Ishii, K. Uehara, Y. Inoko, M. chlorosomes from Chloroflexus aurantiacus. Z. Naturforsch. 46c:228– Kobayashi, and T. Nozawa. 1999. How the formation process influences 232. the structure of BChl c aggregates. Photosynth. Res. 60:229–239. Holzwarth, A. R., and K. Schaffner. 1994. On the structure of Umetsu, M., R. Seki, Z. Y. Wang, I. Kumagai, and T. Nozawa. 2002. bacteriochlorophyll molecular aggregates in the chlorosomes of green Circular and magnetic circular dichroism studies of bacteriochlorophyll c bacteria. A molecular modelling study. Photosynth. Res. 41:225–233. aggregates: T-shaped and antiparallel dimers. J. Phys. Chem. B. 106: Mizoguchi, T., K. Hara, H. Nagae, and Y. Koyama. 2000. Structural 3987–3995. transformation among the aggregate forms of bacteriochlorophyll c as van Rossum, B. J., D. B. Steensgaard, F. M. Mulder, G. J. Boender, K. determined by electronic-absorption and NMR spectroscopies: De- Schaffner, A. R. Holzwarth, and H. M. de Groot. 2001. A refined model pendence on the stereoisomeric configuration and on the bulkiness of of the chlorosomal antennae of the green bacterium Chlorobium tepidum the 8-C side chain. Photochem. Photobiol. 71:596–609. from proton chemical shift constraints obtained with high-field 2-D and Montano, G. A., B. P. Bowen, J. T. LaBelle, N. W. Woodbury, V. B. 3-D MAS NMR dipolar correlation spectroscopy. Biochemistry. Pizziconi, and R. Blankenship. 2003. Characterization of Chlorobium 40:1587–1595. tepidum chloromes: A calculation of bacteriochlorophyll c per chlo- Wahlund, T. M., C. R. Woese, R. W. Castenholz, and M. T. Madigan. rosome and oligomer modeling. Biophys. J. 85:2560–2565. 1991. A thermophilic green sulfur bacterium from new-zealand hot- Nozawa, T., K. Ohtomo, M. Suzuki, H. Nakagawa, Y. Shikama, H. springs, chlorobium-tepidum sp-nov. Arch. Microbiol. 156:81–90. Konami, and Z. Y. Wang. 1994. Structures of chlorosomes and Wang, Z. Y., M. Umetsu, M. Kobayashi, and T. Nozawa. 1999a. Complete aggregated BChl c in Chlorobium tepidum from solid state high assignment of H1 NMR spectra and structural analysis of intact 13 resolution CP/MAS C NMR. Photosynth. Res. 41:211–233. bacteriochlorophyll c dimer in solution. J. Phys. Chem. B. 103:3742– Oelze, J., and J. R. Golecki. 1995. Membranes and chlorosomes of green 3753. bacteria: structure, composition, and development. In Anoxygenic Wang, Z. Y., M. Umetsu, M. Kobayashi, and T. Nozawa. 1999b. C-13- and Photosynthetic Bacteria. R. E. Blankenship, M. T. Madigan, and C. E. N-15-NMR studies on the intact bacteriochlorophyll c dimers in Bauer, editors. Kluwer Academic Publishers, Dordrecht, The Nether- solutions. J. Am. Chem. Soc. 121:9363–9369. lands. 259–278. Wullink, W., and E. F. J. van Bruggen. 1988. Structural studies on Overmann, J., H. Cypionka, and N. Pfennig. 1992. An extremely low-light- chlorosomes from Prosthecochlorisaestuarii. In Green Photosynthetic adapted phototrophic sulfur bacterium from the Black sea. Limnol. Bacteria. J. M. Olson, J. G. Ormerod, J. Amesz, E. Stackebrandt, and Oceanog. 37:150–155. H. G. Tru¨per, editors. Plenum Press, New York. 3–14.

Biophysical Journal 87(2) 1165–1172